mlsanityData to 14 Oct 2026

QArtSR: Quantization via Reverse-Module and Timestep-Retraining in One-Step Diffusion based Image Super-Resolution

arXiv 2503.05584 ↗First posted 7 Mar 2025cs.CVcited by 0

On the map: Diffusion models → Inverse problems and image restoration

Abstract (arXiv)

One-step diffusion-based image super-resolution (OSDSR) models are showing increasingly superior performance nowadays. However, although their denoising steps are reduced to one and they can be quantized to 8-bit to reduce the costs further, there is still significant potential for OSDSR to quantize to lower bits. To explore more possibilities of quantized OSDSR, we propose an efficient method, Quantization via reverse-module and timestep-retraining for OSDSR, named QArtSR. Firstly, we investigate the influence of timestep value on the performance of quantized models. Then, we propose Timestep Retraining Quantization (TRQ) and Reversed Per-module Quantization (RPQ) strategies to calibrate the quantized model. Meanwhile, we adopt the module and image losses to update all quantized modules. We only update the parameters in quantization finetuning components, excluding the original weights. To ensure that all modules are fully finetuned, we add extended end-to-end training after per-module stage. Our 4-bit and 2-bit quantization experimental results indicate that QArtSR obtains superior effects against the recent leading comparison methods. The performance of 4-bit QArtSR is close to the full-precision one. Our code will be released at https://github.com/libozhu03/QArtSR.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Inverse problems and image restoration, one of 14 directions in Diffusion models. Builds on 50 of its 58 references that are in the map. First posted on arXiv 7 Mar 2025.

Builds on

50 of 58 references.

Closest ancestors · 8

Same direction, newest first

  1. 2024Arbitrary-steps Image Super-resolution via Diffusion Inversion12 Dec
  2. 2024PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution26 Nov
  3. 2024Unleashing the Power of One-Step Diffusion based Image Super-Resolution via a Large-Scale Diffusion Discriminator5 Oct
  4. 2024One-Step Effective Diffusion Network for Real-World Image Super-Resolution12 Jun
  5. 2023SeeSR: Towards Semantics-Aware Real-World Image Super-Resolution27 Nov
  6. 2023SinSR: Diffusion-Based Image Super-Resolution in a Single Step23 Nov
  7. 2023DiffBIR: Towards Blind Image Restoration with Generative Diffusion Prior29 Aug
  8. 2023Exploiting Diffusion Prior for Real-World Image Super-Resolution11 May

Other references · 42

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  2. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  3. 2020Denoising Diffusion Implicit Modelscited by 9,781
  4. 2021SwinIR: Image Restoration Using Swin Transformercited by 1,887
  5. 2023Consistency Modelscited by 1,822
  6. 2017Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inferencecited by 1,593
  7. 2018Image Super-Resolution Using Very Deep Residual Channel Attention Networkscited by 1,264
  8. 2014Image Super-Resolution Using Deep Convolutional Networkscited by 1,224
  9. 2023ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillationcited by 1,219
  10. 2021MUSIQ: Multi-scale Image Quality Transformercited by 1,081
  11. 2023One-step Diffusion with Distribution Matching Distillationcited by 1,019
  12. 2022Exploring CLIP for Assessing the Look and Feel of Imagescited by 991
30 more
  1. 2021Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Datacited by 970
  2. 2015Accurate Image Super-Resolution Using Very Deep Convolutional Networkscited by 903
  3. 2020Image Quality Assessment: Unifying Structure and Texture Similaritycited by 811
  4. 2018Residual Dense Network for Image Super-Resolutioncited by 799
  5. 2019Learned Step Size Quantizationcited by 628
  6. 2021Designing a Practical Degradation Model for Deep Blind Image Super-Resolutioncited by 529
  7. 2022ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformerscited by 526
  8. 2022MANIQA: Multi-dimension Attention Network for No-Reference Image Quality Assessmentcited by 522
  9. 2021BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstructioncited by 479
  10. 2023UniPC: A Unified Predictor-Corrector Framework for Fast Sampling of Diffusion Modelscited by 450
  11. 2023OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Modelscited by 424
  12. 2023LLM-QAT: Data-Free Quantization Aware Training for Large Language Modelscited by 324
  13. 2023Q-Diffusion: Quantizing Diffusion Modelscited by 273
  14. 2022Post-training Quantization on Diffusion Modelscited by 267
  15. 2023LoftQ: LoRA-Fine-Tuning-Aware Quantization for Large Language Modelscited by 196
  16. 2019Low-bit Quantization of Neural Networks for Efficient Inferencecited by 193
  17. 2020LSQ+: Improving low-bit quantization through learnable offsets and better initializationcited by 189
  18. 2023PTQD: Accurate Post-Training Quantization for Diffusion Modelscited by 170
  19. 2024SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Modelscited by 170
  20. 2023Dual Aggregation Transformer for Image Super-Resolutioncited by 124
  21. 2022Cross Aggregation Transformer for Image Restorationcited by 92
  22. 2024BitDistiller: Unleashing the Potential of Sub-4-Bit LLMs via Self-Distillationcited by 84
  23. 2021Nonuniform-to-Uniform Quantization: Towards Accurate Quantization via Generalized Straight-Through Estimationcited by 79
  24. 2023EfficientDM: Efficient Quantization-Aware Fine-Tuning of Low-Bit Diffusion Modelscited by 78
  25. 2023Towards Accurate Post-Training Quantization for Vision Transformercited by 71
  26. 2024QuEST: Low-bit Diffusion Model Quantization via Efficient Selective Finetuningcited by 49
  27. 2024Low-Rank Quantization-Aware Training for LLMscited by 43
  28. 2024BitsFusion: 1.99 bits Weight Quantization of Diffusion Modelcited by 37
  29. 2024Improving Quantization-aware Training of Low-Precision Network via Block Replacement on Full-Precision Counterpartcited by 6
  30. 2023SmoothQuant+: Accurate and Efficient 4-bit Post-Training WeightQuantization for LLMcited by 5

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Q-DiT4SR: Exploration of Detail-Preserving Diffusion Transformer Quantization for Real-World Image Super-Resolution
  2. 2025QuantVSR: Low-Bit Post-Training Quantization for Real-World Video Super-Resolution
  3. 2026LSGQuant: Layer-Sensitivity Guided Quantization for One-Step Diffusion Real-World Video Super-Resolution
  4. 2026QuantSR+: Pushing the Limit of Quantized Image Super-Resolution Networks
  5. 2026QUSR: Quality-Aware and Uncertainty-Guided Image Super-Resolution Diffusion Model
  6. 2025Q-Sched: Pushing the Boundaries of Few-Step Diffusion Models with Quantization-Aware Scheduling
  7. 2026TEASR: Training-Efficient Any-Step Diffusion Transformer for Real-World Image Super-Resolution
  8. 2023Post-training Quantization for Text-to-Image Diffusion Models with Progressive Calibration and Activation Relaxing